Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Random-effects models for analyzing clustered data from a nutrition education intervention.

S I Woodruff1

  • 1San Diego State University, USA.

Evaluation Review
|November 3, 1997
PubMed
Summary

Clustered data in public health interventions can skew results. Random-effects models (REMs) analyzing subjects within classrooms or sites offer more accurate intervention effect estimates than standard regression.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Self-selection in responding to a health risk appraisal: are we preaching to the choir?

American journal of health promotion : AJHP·2011
Same author

Intervention to reduce environmental tobacco smoke exposure in Latino children: null effects on hair biomarkers and parent reports.

Tobacco control·2004
Same author

Evaluation of a culturally appropriate smoking cessation intervention for Latinos.

Tobacco control·2002
Same author

On-line smoking cessation: what's porn got to do with it?

Tobacco control·2002
Same author

Pilot test of an Internet virtual world chat room for rural teen smokers.

The Journal of adolescent health : official publication of the Society for Adolescent Medicine·2001
Same author

Compliance with federal and state legislation by indoor tanning facilities in San Diego.

Journal of the American Academy of Dermatology·2001

Area of Science:

  • Public Health
  • Biostatistics
  • Nutrition Science

Background:

  • Public health interventions often involve group-level randomization and implementation.
  • Ignoring data dependency from clustering can lead to inaccurate intervention effect estimations.
  • Standard regression analyses may not adequately account for hierarchical data structures.

Purpose of the Study:

  • To compare the efficacy of different statistical methods for analyzing clustered data in public health interventions.
  • To evaluate the performance of random-effects models (REMs) against traditional regression approaches.
  • To determine the most appropriate analytical strategy for nutrition intervention data with nested subjects.

Main Methods:

  • Analysis of nutrition intervention data using four distinct statistical approaches.

Related Experiment Videos

  • Method 1: Standard multiple regression with individual subject data.
  • Method 2: Standard multiple regression using the classroom as the unit of analysis.
  • Method 3: Two-level random-effects model with subjects nested within classrooms.
  • Method 4: Two-level random-effects model with subjects nested within sites.

    • Main Results:

    • Standard regression using individual data may overlook clustering effects.
    • Regression using the cluster as the sole unit of analysis can reduce statistical power and obscure individual-level effects.
    • Two-level random-effects models provide a more nuanced analysis by accounting for both individual and cluster levels.
    • The choice of clustering unit (classroom vs. site) in REMs impacts the interpretation of intervention effects.

    Conclusions:

    • Random-effects models are recommended for analyzing public health intervention data with clustered structures.
    • Accurate analysis requires accounting for the hierarchical nature of data, distinguishing between individual and cluster levels.
    • The findings highlight the importance of selecting appropriate statistical methodologies to avoid misrepresenting intervention effects in public health research.